12 Simulations, Requirements, and Detector Performancehep.ucsb.edu/LZ/TDR/12_Simulations_Requirements_and... · 2017-03-26 · Figure 12.1.1: Engineering drawing and simulation geometry

12 Simulations, Requirements, and DetectorPerformance

This chapter ties together the previous descriptive chapters and describes the performance we expect fromthe LZ apparatus. Our estimates of performance are based on detailed simulations, which have been sub-stantially overhauled since completion of the LZ Conceptual Design Report [1]. The LUX collaboration hasperformed extensive high-statistics calibrations of ER background [2], NR signal [3], and implemented thesecalibrations in an improved analysis using the profile likelihood ratio technique [4] of the LUX WIMP searchdata [5].

We have incorporated the performance improvements made by LUX, which change the sensitivity in twoprincipal manners:

1. The LZ sensitivity to low-mass WIMPs and to nuclear recoils from solar 8B neutrinos is much im-proved.

2. The LZ sensitivity to sources of ER background, including, prominently, the "quiet" beta decay ofdaughters that trace their parentage to radon impurities, is much reduced.

12.1 Simulations

In this section we describe in detail the simulations used to develop the background and sensitivity studiespresented in this report. The simulations were performed using LZSim, an offshoot of the LUXSim pack-age originally developed for the LUX experiment and based on the GEANT4 particle physics simulationsoftware [6, 7]. Designed specifically for low-background detector modeling, LZSim generates events andrecords particle interactions on a detector geometry component-by-component basis, but with an infrastruc-ture independent of the detector geometry.

The LZSim infrastructure allows the user to define any detector component as a GEANT4 sensitive detectorat run-time with a macro command. It incorporates internal bookkeeping to automate the generation ofbackgrounds arising from a variety of event generators, each with intensities set by the user at run time,and with a time-ordered stochastic primary event record. The geometry components are customized usingcoding techniques familiar to users of the base GEANT4 code. The event generators can be based eitheron the internal GEANT4 classes or created from scratch by the users. LZSim automatically records qualitycontrol information to a header in each output file to establish a record of how the data was generated.

We use GEANT4 version 4.9.5, the physics list QGSP_BIC_HP, and the libraries of CLHEP version 2.1.0.1.LZSim provides the option to incorporate the NEST model that describes ionization and scintillation forma-tion for NRs and ERs [8, 9]. Currently we instead pass the output from LZSim to a standalone version of theNEST model that incorporates the latest results from the LUX experiment [2, 5].

12 Simulations, Requirements, and Detector Performance LZ Technical Design Report

12.1.1 Geometry construction

A detailed model of the LZ detector geometry was created within the LZSim framework according to CADdrawings of the detector. Components with significant mass or high amounts of radio-impurities as well asthose located very close to the active xenon target are included. Elements that influence light collection andtheir respective optical properties are also described in the model.

Our model describes the detector from the outside in, with a few exceptions, nesting each successive vol-ume or component within the one preceding it. The outermost volume is the steel water shielding tank.Placed within the water of the shielding tank are the major components of the outer detector of Chapter 4,including the segmented acrylic vessels, liquid scintillator, foam displacer, reflectors and R5912 PMTs. Ser-vices for the TPC such as the cryostat support stand, cathode high voltage conduit and thermosyphon andPMT cabling conduits require penetrations in the acrylic tanks that can impact veto performance and aretherefore implemented in the detector model. Conduits that contain multiple or complex materials such ascoaxial cable, gaseous or liquid xenon, and vacuum are modeled by a single material which represents theaverage density and composition of all materials in that conduit. Angled and horizontal neutron calibrationtubes and a port for the YBe source are included for calibration source studies.

Located within the outer detector is the titanium cryostat, built according to the engineering models, asshown in Fig. 12.1.1. An inner cryostat vessel is contained within the outer cryostat vessel, also made oftitanium and built to engineering specifications. The model includes multi-layer insulation between thesetwo volumes, in addition to vacuum. No optical properties are defined between these regions as no photonsare expected to be produced here.

P. Beltrame

Components with major impact on background frozen on Oct. 20th

Inner and Outer vessels Stainless Steel rings around vessels

PMT arrays PMT geometry including bases PMT support truss

PTFE walls in Liquid and in Gas Grid rings in PTFE Grid - cross wires Field rings metal (including Reverse Field Region) HV resistors

13

Scott Haselschwardt, Jeremy Mock, Slava Bygayov, Eric Miller, Elena Korolkova, Sally Shaw, Ty Stiegler…

Figure 12.1.1: Engineering drawing and simulation geometryof the outer cryostat.

The first liquid xenon volume,known as the xenon skin, is just insidethe inner cryostat vessel and continuesinward to the outer surface of theTPC. The skin also extends below thebottom of the TPC through the bottomdome (see Chapter 3 for more details).The inner wall of the cryostat in thesimulation is covered in a thin PTFEliner to aid light collection.

Optical boundaries, which can be de-fined at run time, are used between theliquid xenon skin and the componentswithin to study the effects of changes inthe assumed reflectivities.

The heart of the experiment, the liq-uid xenon TPC where the primary scin-tillation (S1) and secondary scintilla-

tion from ionization (S2) signals would be caused by a WIMP, is nested inside the skin volume. A renderingof the TPC is shown Fig. 12.1.2. The liquid xenon in the TPC is divided into two volumes. The active volumecontains all the liquid xenon above the cathode, where well-reconstructed S1/S2 events occur. The secondvolume is the reverse field region (RFR), below the TPC cathode. Energy depositions in this volume causean S1 signal, but no S2. Nevertheless, sometimes an RFR S1 signal becomes associated with an unrelatedS2 signal, resulting in a class of events known as “gamma-X”. A third xenon volume, the gaseous xenonabove the liquid where electroluminescence develops for the S2 signal, is also defined.

A cylinder of PTFE with extremely high diffuse reflectivity forms the TPC wall. Field shaping rings andgrading resistors in the walls are represented. TPC components including the bottom shield grid, cathode

320

LZ Technical Design Report 12.1 Simulations

grid, gate grid, and anode grid, according to the design described in Table 3.6.3, are represented. Opticalboundaries between the liquid xenon and the PTFE walls as well as between the liquid xenon and the stainlesssteel of the grid wires are included to describe the light collection properties of the detector.

P. Beltrame





13


Figure 12.1.2: Engineering drawing and simulation geometryof the TPC

At the top and bottom end of the TPCcylinder are the PMT arrays, containingrepresentations of each R11410 PMTthat will be used in the LZ experiment.The PMTs are described as a stainlesssteel shell with vacuum inside. Mate-rials representing the dynode chain areincluded to model the radioactivity ofthose components.

The face of the PMT is a quartz win-dow with an additional thin layer ofquartz buried within to represent thephotocathode. The window and photo-cathode volumes are divided to modelcases where a photon reflects off thefront face of the PMT and fails to pen-etrate to the photocathode.

Optical interfaces at the top of the TPC between the PMT windows and gaseous xenon in the extractionregion, and at the bottom array between the PMT windows in liquid xenon, are described so GEANT4 appro-priately handles the reflection and transmission of photons. The PMT photocathode is defined as a GEANT4sensitive detector to collect optical photons. The PMT array volumes also contain the support plates and sup-port trusses that provide the mechanical support for the PMTs. Figure 12.1.3 shows a comparison betweenthe current engineering design and the simulated geometry of the bottom PMT array.

P. Beltrame





13


Figure 12.1.3: Engineering drawing and simulation geometry of the bottom PMT array and supporttruss. The skin PMTs that mount to this truss are not shown.

12.1.2 Event Generators

A variety of software packages are employed to simulate the physics of signals and backgrounds that induceresponses in the LZ detector. The event generators which describe the WIMP signal and neutrino physics arederived from the references cited in Chapter 2. Here we discuss the event generators employed to describevarious types of background phenomena in the LZ detectors.

321


12.1.2.1 Neutron production in detector materials

Neutrons emitted from radioactive processes in the material near to the LZ liquid xenon detector can createisolated nuclear recoils that might fake the recoils expected from WIMPs. To simulate neutron backgroundsfrom radioactivity (the 238U, 235U and 230Th decay chains), LZSim uses input neutron spectra as calculatedwith the SOURCES-4A [10] package.

The SOURCES-4A code calculates neutron yields and spectra from spontaneous fission, (� ,n) reactionsand delayed neutron emission due to the decay of radionuclides. Its library contains all alpha emission linesfrom known radioactive isotopes. The code takes into account the energy losses of alphas, cross-sections of(� ,n) reactions and the probabilities of nuclear transition to different excited states. We use an option for athick target neutron yield allowing for calculation of neutron yields and spectra under the assumption thatthe size of a material sample exceeds significantly the range of alphas.

The original SOURCES-4A code has been modified [11–13] to extend the energy range of alpha particlesto 10 MeV and to include (� ,n) cross-sections and transition probabilities to excited states for most isotopesrelevant to underground rare event experiments, based either on measurements or on EMPIRE1.19 [14].

The neutron spectra from SOURCES-4A are implemented as generators in LZSim, allowing any detectorcomponent to become a source of neutrons. For each component of interest, we simulate concentrationsof 10 ppb of uranium or 10 ppb of thorium; these concentrations are then scaled to match the results of thematerials screening. The uranium decay chains are split into early and late branches, and the 210Pb sub-chainis calculated separately.

The spontaneous fission (s.f.) process is treated separately to exploit the ability of LZ to reject decaysproducing multiple neutrons and gammas, such as those produced in s.f. events. The SOURCES-4A packagegenerates individual neutron spectra without accounting for simultaneous multiple neutrons and gamma rays.A special generator was developed to accurately simulate multiple neutrons (2.01 on average for 238U) andgammas (6.44 on average for 238U) in s.f. events. Accounting for multiple neutrons and gammas permitsaccurate description of multiple simultaneous signals in the LZ detector, and accurate accounting of therejection of events with multiple signals. Most spontaneous fission events, particularly those in materialsclose to the LXe target, are rejected through their tendency to produce multiple hits.

12.1.2.2 Muon-induced neutrons

Energetic neutrons can be produced by atmospheric muons that penetrate to the rock around the Davis Cavern.Simulations of this process commence with the selection of muons with positions, directions, and energiessampled according to the MUSUN code [15] after transport using standalone MUSIC code [16]. MUSUNhas been integrated into LZSim as a particle generator in such a way that events are generated from thesurface of a cuboid surrounding the detector.

Muons sampled with the MUSUN code are then passed to GEANT4 which transports them and their sec-ondaries including neutrons to the detector. All processes relevant to muon, photon, electron, and hadroninteractions are included and the models are equivalent to those used in the GEANT4 physics list calledShielding, recommended by the GEANT4 developers for this application.

12.1.2.3 Gamma activity in large detector components

Background ER events from 40K, 60Co, and the 238U and 232Th decay chains are also modeled using aparticle generator developed in LZSim. Ions of the parent isotopes are positioned in a component and the fulldecay chain is simulated according to the physics described by the GEANT4 radioactive decay data libraries.Crucially, the decay chain is produced in equilibrium, which allows splitting of the chain by individualisotopes during analysis. Thus, only one simulation is required, independent of the relative decay rates of

322


individual radioactive isotopes within the chain. As for the neutron studies, fixed activities are simulated foreach component, which are then renormalized based on materials screening results.

12.1.2.4 Gammas from the cavern rock

The external background from gammas in the rock walls of the cavern are also assessed using the radioactivedecay chain generators described in the previous section. However, due to the large number of primarydecays required to accumulate events in the active xenon, event biasing is used to boost statistics. Thisinvolves saving events at decreasing distances from the target and feeding them back into LZSim multipletimes as primary particles. By default, LZSim records all hits within a detector component with a definedrecord level, but for these simulations the structure has been modified so that the output is recorded for tracksaccording to the distance from the center of the TPC. Additionally, modifications were made to allow theLZSim binary output files to be input as primary particles.

12.1.2.5 Benchmark points

A new generator has been developed to allow the assessment of the impact of the radioactivity of small sizecomponents (e.g. sensors) quickly. LZSim was designed to primarily distribute primary events throughouta given volume. Our new generator can associate radioactive events with a specific geometric location (the‘Benchmark Point’) within the LZSim geometry.

12.1.3 From energy deposition to signals

The event generators described in the previous section are used to generate different radioactive decay prod-ucts. These particles are tracked by GEANT4 as they deposit their energy in different volumes of the geometry.LZSim allows any part of the volume to be a sensitive detector, and for these simulations, we record all depo-sitions in the TPC, as well as associated energy depositions in the skin region and the outer detector. Amongthe data saved in each volume for each event are the locations, times, and magnitudes of energy depositsmade by various particle types.

In the skin and TPC liquid xenon volumes, energy depositions within 400 µm are clustered (to encompassthe largest possible ER and NR track sizes) and categorized as either ER or NR depending on the interaction.The result is a list of energy-deposition clusters from a given particle type (ER or NR). The energy andassociated particle type of each cluster as well as the local electric field are then fed into the NEST (NobleElement Simulation Technique) [8, 9, 17] package which stochastically computes the number of expectedphotons and electrons produced at each cluster.

NEST models the scintillation light and ionization charge yields of nuclear and electron recoils as a func-tion of electric field and energy or dE∕dx. NEST also models the drift, diffusion, absorption, extraction,and electroluminescence of the electrons as they move through the liquid and gas. “NEST” refers both to acollection of microscopic models for energy deposition in noble elements and to the Monte Carlo simulationcode that implements these models. NEST provides mean yields and intrinsic fluctuations due to the physicsof excitation, ionization, and recombination, including both Gaussian and non-Gaussian components of theenergy resolution.

The NEST methodology was initially trained on data from a small dual-phase detector from Case West-ern Reserve University (Xed), which yielded comprehensive data sets in terms of energy range and fieldsweep [18]. The NEST model used in this simulation has been updated to incorporate the latest calibrationresults from the LUX experiment [2, 3, 5].

323


After the clustered energy depositions have been converted from energy to quanta (scintillation photonsand ionization electrons) via the NEST models, a detector model is applied to convert these raw quanta intothe detector observables, that is, into S1 and S2.

The primary scintillation light from the NEST models is propagated to the faces of the PMTs, accountingfor binomial fluctuations, using the light collection model, including Fresnel transmission and reflection,described in Section 3.5. We refer to the average light collection efficiency as �1, which the optical modelcurrently predicts to be 8.5 % (better than our baseline value and requirement of 7.5 %), and we correctthe raw signal (denoted S1) for the variation of light collection with position in the detector (denoted S2).Photoelectron production at the photocathode accounts for the double-photoelectron phenomenon describedin [19]. A trigger requirement of three-fold coincidence of PMT hits is applied prior to subsequent analysis.Both S1 and S1c are reported to the user.

The S2 signal is produced from the number of ionization electrons that drift away from the interactionsite. The LZ extraction efficiency (see Table 3.6.1) is applied to determine how many electrons are extractedto the gas phase. The extracted electrons are converted to a number of luminescent photons with NEST,accounting for LZ parameters, and these photons are then converted into photoelectrons at the PMTs. TheS2 signal is also corrected for the position of the event in the detector, including the effect of non-infiniteelectron lifetime presumed in LZ to be 850 µs (corresponding to an absorption length of greater than 1.5 m).The raw and corrected signals, denoted S2 and S2c, are reported to the user.

In the central TPC the energy-weighted position and variance of all energy clusters is computed. The totalS1 signal in the central xenon volume includes the S1 signal created by energy deposits in the reverse fieldregion (RFR) below the cathode. Here, the electric field drifts electrons downward, making no contributionto the S2 signal.

A similar procedure is followed to model the S1 signal observed in the skin region (excluding S2 as nocharge is collected in the skin). A light collection model is determined by simulating photons throughoutthe skin. Each energy deposit in the skin is converted into photon quanta using an electric field model of theskin and the NEST package. These photons are then converted to detected photoelectrons, and the final skinS1 is reported to the user.

The process that translates energy depositions into observed S1c and S2c signals does not currently accountfor PMT or DAQ electronics noise. A more complete model incorporating all known sources of noise leadingto the generation of simulated waveforms is under development.

A substantial effort has gone into simulating the light collection for the outer detector liquid scintillatorsystem. Currently, however, for the primary TDR physics simulations, only energy depositions from GEANT4in the outer detector are used. The transport, capture, and conversion to gamma rays and nuclear recoils ofneutrons are simulated in GEANT4.

12.1.4 Analysis cuts

For each event, we have a tree containing the energy and location of interactions in the TPC, the skin region,and the outer detector, and energy depositions in liquid xenon have been translated into raw and correctedS1 and S2 as described in the previous section. We apply a set of cuts to the simulation data to determinethe backgrounds produced by a given radioactive decay chain in a given detector component. The baselinecuts that are used to produce the numbers in Table 9.2.7 are as follows:

• Region of Interest: 0 < S1c < 20 detected photoelectrons, but assuming 3-fold coincidence in theTPC PMTs. In other words, three PMTs have to have observed light, but the total sum of the signalcan be arbitrarily small. For ER, this range is approximately 1.5 to 6.5 keVee, and for NR, this rangeis approximately 6 to 30 keVnr. In addition, the uncorrected S2 signal is required to be >350 detectedphotoelectrons (5 emitted electrons) ensuring adequate signal size for position reconstruction.

324


• Single scatter event: �Z < 0.2 cm and �r < 3.0 cm. The energy-weighted variance in position must beless than the expected spatial resolution of the detector for an event to be classified as a single scatteringevent. The given values are based on the LUX performance with an estimated scaling to LZ. Gamma-Xevents are treated as single scatters, as no S2 can be observed from the second interaction vertex.

• Skin cut:


12.2 Requirements

In this section, we summarize the key requirements for the LZ experiment. The LZ collaboration has es-tablished a small number of requirements to guide and evaluate the design and fabrication of the detectorsystems.

The top-level scientific requirement is the sensitivity to WIMP dark matter, via the spin-independent scat-tering process. Subsidiary high-level science requirements and the flow-down from the overall sensitivityare shown in Fig. 12.2.1. The high-level requirements, including the key infrastructure requirements, aresummarized in Table 12.2.1. These requirements flow down to the detector subsystems and are capturedin a concise form available to the collaboration. There are two practical high-level requirements. First, allequipment and subassemblies must be transported via the Yates shaft (see Chapter 10), which imposes di-mensional and weight limits. Second, the existing water tank now housing the LUX detector will be reused.The collaboration has also captured the requirements for detector subsystems at WBS Level 2.

Sensi&vityR-0001

FiducialExposureR-0002

AnalysisThresholdR-0003

Discrimina&onR-0004

InternalBackgroundsR-0005

Ac&veMassR-0006

ExternalBackgroundsR-0007

Singlee-Detec&onR-0008

SinglepheDetec&onR-0009

S1Collec&onR-0010

Level2Requirements

1

Figure 12.2.1: Flow down of the top level scientific requirements.

Requirements development and explication have been key elements of internal reviews of LZ detectorsystems and will be an important aspect of configuration control. All top-level and Level-2 requirementshave been developed and reviewed in dedicated meetings. The requirements are captured in a Google doc-ument, along with additional material and documentation for each relevant WBS Level-2 element. The LZinstrument scientist and chief and deputy chief engineers (see Chapter 13) are primarily responsible for themaintenance of the requirements and their further development, if required, in close collaboration with theLevel 2 managers.

326

LZ Technical Design Report 12.2 Requirements

Table 12.2.1: LZ Top Level Science Requirements

Number WBSRequirementName

Requirement Descrip-tion

Rationale

Primary

R-0001 ScienceWIMP Sensitiv-ity

Sensitivity to40GeV/c2 WIMPsis 3 × 10−48 cm2 orbetter

Probe limit of liquid xenon tech-nology set by solar neutrino back-ground. Approach sensitivity toatmospheric neutrinos. Test su-persymmetric and extra-dimensionmodels of dark matter

Secondary

R-0002 ScienceFiducial Expo-sure

Minimum of5,600 tonne-days

Needed to achieve sensitivity re-quirement. Achievable with fiducialmass of 5.6 tonnes and assumedrunning period of 1,000 live days orless fiducial mass and longer run-ning time

R-0003 ScienceAnalysis thresh-old

50% efficiency at6 keVnr

Probe WIMP mass range down to5GeV/c2 with non-negligible sensi-tivity

R-0004 ScienceER Discrimina-tion

99.5% ER discrimi-nation for 50% NRacceptance

Limit background from ERs so asto reach WIMP sensitivity require-ment

R-0005 ScienceInternal back-grounds

Internal backgroundsfrom radioactive no-ble gases (Rn, Kr, Ar)not to exceed fourtimes the solar neu-trino ER background

Limit ERs from internal back-grounds to an acceptable level. So-lar neutrino rate does not include8B

TertiaryR-0006 Science Active mass 7.0 tonnes Required to reach fiducial exposure

R-0007 ScienceExternal back-grounds

Backgrounds fromradioactivity of thedetector components(not including in-ternal backgrounds,R-0005). ER countsbefore discrimination


Table 12.2.1: (continued)


Requirement Descrip-tion

Rationale

R-0008 ScienceSingle electrondetection

50 S2 photoelectronsdetected per emittedelectron

Sufficiently large S2 signal for ac-curate reconstruction of peripheralinteractions, such as those arisingfrom contamination on the TPCwalls.

R-0009 ScienceSingle pho-toelectrondetection

Single S1 photoelec-tron detection with>90% efficiency, soas to reach >70% ef-ficiency for 3 phe

Main determinant of analysisthreshold.

R-0010 ScienceS1 light collec-tion

Volume-averaged S1photon detectionefficiency (geometriclight collection timeseffective PMT quan-tum efficiency) of>7.5%

Good discrimination and low en-ergy threshold, equal or better thanpast Xe experiments. Exponentiallyfalling (in recoil energy) WIMPspectrum means more recoils atlower energies, and low-energy re-coils produce less S1 (both totaland per-unit-energy) driving the S1light collection efficiency require-ment.

Infrastructure

R-0100 GeneralAll parts fitdown Yatesshaft

All detector elementsmust be sized so thatthey can be loweredvia the Yates shaft

Yates shaft is primary access to theDavis campus

R-0110 GeneralReuse Davis wa-ter tank

Existing Davis watertank is reused. In-clude minor modifi-cations and refurbish-ment.

Not practical or cost effective toreplace water tank. Insufficientunderground space to make largertank.

The baseline design described in this Technical Design Report meets or exceeds all requirements. Webriefly summarize the key requirements by WBS element in the subsections below. Linkage among thevarious requirements has been part of the requirements review process.

328


12.2.1 WBS 1.1 Xenon Procurement

There is a single requirement for WBS 1.1 Xenon Procurement, that being to procure 10 total tonnes ofxenon. This requirement flows down from R-0006, as 10 total tonnes are adequate to result in 7 active tonnesin the detector.

329


12.2.2 WBS 1.2 Xenon Vessel

The primary requirements for the cryostat vessels are shown in Table 12.2.2. The size of the cryostat deter-mines the amount of target that can be deployed and flows down from the top level science requirement butalso the infrastructure requirement that all parts must fit in the Yates shaft. The cryostat vessels must alsobe low radioactivity. Safety considerations dictate that the vessels are compliant with standard engineeringcodes.

Table 12.2.2: Level 2 Requirements for WBS 1.2 Xenon VesselNumber WBS Requirement Name Requirement Description Rationale

R-120001 1.2Inner Cryostat Ves-sel (ICV) MaximumOuter Diameter

Inner cryostat outside di-ameter less than 1.702 m(67.0 inch). Ports must beat angles where there is suf-ficient clearance.

Must fit within Yatesshaft.

R-120002 1.2Inner cryostat ves-sel (ICV) compactgeometry

Design compact ICV to min-imize use of passive xenon:ICV conical section and el-lipsoidal 3:1 dished end

Saving xenon

R-120003 1.2Outer CryostatVessel segmented

No fabrication undergroundSegmented OCV fits intoYates shaft

R-120004 1.2Low radioactivity ofthe cryostat mate-rial

Radioactivity budget frombackground simulations.Maximum contribution tothe overall background:3.3% of pp solar neutrinosand 0.03 NR event

Low radioactivity, lowdensity minimizing, effi-ciency of Outer Detector

R-120005 1.2

Outer CryostatVessel (OCV) andInner Cryostat Ves-sel (ICV) designedfor 1.48 bar exter-nal pressure andvacuum internal

Working conditions for OCVand most severe failuremode for ICV

Vacuum inside and wateroutside

R-120006 1.2

Inner Cryostat Ves-sel (ICV) designedfor 4 bar inner pres-sure and vacuumexternal

ICV working conditions

Xe gas inside with maxi-mum pressure - includinghydrostatic pressure atthe bottom dished end;3.4 bar top head

R-120007 1.2Design complianceto codes

Compliance to ASME BPVCcode VIII, 2012 Int. BuildingCode, ASCE 7 with soil clas-sification Cals B for seismicconditions, Fabricator holdsU-stamp certificate

Required by SURF SDregulations

330


12.2.3 WBS 1.3 Cryogenic System

The primary requirements for the Cryogenic System are shown in Table 12.2.3. The main objective of thissystem is to enable safe and efficient cooling of the LXe target volume. In particular, the cooling power mustbe adequate to enable rapid circulation of the Xe volume for purification to satisfy the Level-2 requirementsin WBS 1.4, which is an example of linkage between WBS requirements.

Table 12.2.3: Level 2 Requirements for WBS 1.3 Cryogenic System


Requirement Description Rationale

R-130001 1.3Sufficient cool-ing power

Cryogen cooling systems shallbe sufficient to remove heat for500 slpm of Xenon circulation;purge of the detector Xenon gasspace; and thermal losses of allthe system components.

There must be adequatecooling to liquefy Xenonwithin the parameters offlow required to attainpurity.

R-130002 1.3Oxygen defi-ciency hazardsafety

Engineered controls shall beimplemented to achieve ODHClass 2 or better.

SURF requires ODHClass 2 or better for allexperimental implemen-tations

R-130003 1.3Pressure safety- valves

Redundant relief devices shallbe employed for pressure safety.Relief devices shall be sized perCGA S-1.3.

Properly sized redundantrelief devices are a pri-mary engineered controlfor pressure systems.

R-130004 1.3Pressure safety- monitoring

Active pressure monitoring shallbe utilized such that alarms pro-vide warning of pressure goinghigher than the planned operat-ing pressure range.

Monitoring of pressuremay provide an opportu-nity to act on elevatedpressure before one ofthe primary safety de-vices is activated.

R-130005 1.3 Materials

Materials exposed to LN tem-peratures shall be: a) low car-bon stainless steel; b) aluminumbased alloys; c) nickel based al-loys; d) copper / copper basedalloys; or e) pure titanium.

These materials will re-main ductile at 77 K.

R-130006 1.3Thermal insula-tion

Equipment at 175K shall havea minimum of 10 layers andequipment at 77 K shall have aminimum of 25 layers of multi-layer insulation (MLI).

MLI is the most effec-tive method of minimiz-ing radiative heat load.Lower operating temper-atures require additionallayers of MLI.

331


12.2.4 WBS 1.4 Xenon Purification

The primary requirements for WBS 1.4 Xenon Purification are shown in Table 12.2.4. These requirementsprimarily flow down from R-0003 and R-0005, as the Xe must be pure enough to enable efficient extractionof signals to satisfy the analysis threshold requirement but also low enough in internal radioactive sourceslike Kr and Rn to satisfy the internal backgrounds requirement.

Table 12.2.4: Level 2 requirements for WBS 1.4 Xenon PurificationNumber WBS Requirement Name Requirement Description Rationale

R-140001 1.4Xe electronegativepurity

Charge absorption length>1.5m, O2 equivalent=0.4 ppb.

Collect charge and scin-tillation throughout theentire volume of theTPC.

R-140002 1.4Removal of Kr andAr from Xe.

Ability to remove natural Krfrom the Xe to a concentra-tion of


12.2.5 WBS 1.5 Xenon Detector System

There are several requirements in WBS 1.5 Xenon Detector System. For example, the TPC must be largeenough to accommodate the required target mass, the electric fields need to be adequate to achieve effi-cient single electron detection and to achieve the required 99.5 % ER/NR discrimination, the TPC must haveadequate light collection, and the detector components must be made from clean materials to limit the ex-ternal backgrounds seen by the LXe. R-150009 is another good example of the linkages between differentWBS systems, as the power dissipated in the detector must be low enough that the cooling power speci-fied in R-130001 is adequate. The primary requirements for WBS 1.5 Xenon Detector System are listed inTable 12.2.5.

Table 12.2.5: Level 2 Requirements for WBS 1.5 Xenon Detector System

Number WBS Requirement NameRequirementDescription

Rationale

R-150001 1.5TPC inner dimen-sions

Dia. =1,456mm /length =1,456mm

7 tonne active mass, optimal self-shielding

R-150002 1.5 TPC drift field 300V/cm99.5% discrimination; drift time95% ofskin volume

Veto efficiency required for fiducialmass

R-150006 1.5Component ra-dioactivities

<250∕240∕240∕540mBq U/Th/Co/K


12.2.6 WBS 1.6 Outer Detector

WBS 1.6, the Outer Detector, acts as the main neutron veto, and therefore most requirements in this WBSflow down from R-0007. The Level 2 Requirements for WBS 1.6 Outer Detector are shown in Table 12.2.6.

Table 12.2.6: Level 2 Requirements for WBS 1.6 Outer DetectorNumber WBS Requirement Name Requirement Description Rationale

R-160001 1.6Neutron veto effi-ciency

Detection efficiency of 95 %for a 1 MeV neutron thatscatters once in the xenon

Needed to meet the neu-tron part of requirementon external backgrounds

R-160002 1.6Outer Detectorthreshold

Threshold of 200 keV (50 %efficiency) on energy depositin liquid scintillator

Highest achievable vetoefficiency with accept-able deadtime

R-160003 1.6Number of photo-electrons detectedper energy deposit

>80 phe/MeVCorresponds to about15 phe at threshold of200 keV

R-160004 1.6Light collection ef-ficiency

Efficiency of >5% for havingVUV photons strike a PMT

Light detection is pro-portional to light collec-tion efficiency.

R-160005 1.6 DeadtimeOD must not veto more than5% of the WIMP search live-time

Keep overall activity lowto limit deadtime im-posed by OD veto trig-gers

12.2.7 WBS 1.7 Calibration System

The Calibration System must provide accurate calibrations of the light and charge yields, position dependen-cies, time variations, energy threshold, resolution, and discrimination parameters of the central TPC withouttaking an undue amount of running time away from the primary dark matter search. The Calibration Systemmust also provide an accurate picture of the performance of the veto systems: the Xe skin and Outer Detector.The L2 requirements for WBS 1.7 Calibration System are listed in Table 12.2.7.

Table 12.2.7: Level 2 Requirements for WBS 1.7 Calibration SystemNumber WBS Requirement Name Requirement Description Rationale

R-170001 1.7 Calibration Times


Table 12.2.7: (continued)Number WBS Requirement Name Requirement Description Rationale

R-170002 1.7Calibrate TPC(x,y,z) variation


12.2.8 WBS 1.8 Electronics, DAQ, Controls, and Computing

The primary requirements for WBS 1.8 Electronics, DAQ, Controls, and Computing are shown in Ta-ble 12.2.8. The main requirement here is to enable a low energy threshold in the detector. However, thereare two requirements that flow across from WBS 1.7, as the data acquisition must be robust enough to han-dle the event rates listed in Table 12.2.7 needed to effectively calibrate the detector. On the controls side,this subsystem is responsible for providing monitoring and control during emergencies, and in particular thecontrol of the xenon recovery system of WBS 1.4.

Table 12.2.8: Level 2 Requirements for WBS 1.8 Electronics, DAQ, Controls, and Computing


Rationale

R-180001 1.8 Energy threshold90% efficiency fora single phe

The S1 analysis threshold is de-fined by our efficiency to capturesingle-photoelectron signals. Fora specific gain, this requirementfixes the noise requirements of theelectronic chain. This calcula-tion depends sensitively on the as-sumed 35 % variations in the PMTresponse.

R-180002 1.8 Source count rate 150HzAble to handle source calibrations.Rates are limited by the drift time.

R-180003 1.8 LED count rate 4 kHzAble to handle LED calibrationrates.

R-180004 1.8

Guarantee thesafety of xenonsupply and thexenon circulationsystem

Use ProgrammableLogic Controllers(PLCs) to controland monitor thexenon purification,circulation, andstorage systems

PLCs are used to ensure that thedetector and xenon system aremaintained in a safe state duringemergencies that result in a shut-down of the slow-control infras-tructure. The PLC system pro-vides continuous real-time moni-toring and control of the criticalsubsystems and will initiate auto-matic recovery of the xenon tothe storage facility during an emer-gency.

336


12.2.9 WBS 1.9 Integration and Installation

The primary requirements for WBS 1.9 Integration and Installation are shown in Table 12.2.9. There are twomain requirements in this WBS. The first is that all parts must be installed correctly, and the second is thatthe detector must be installed with clean parts in such a way to minimize exposure to radon and dust. Theserequirements flow down from the top level science requirements, particularly R-0005 and R-0007.

Table 12.2.9: Level 2 Requirements for WBS 1.9 Integration and Installation


Requirement Description Rationale

R-190001 1.9Parts usedhave acceptableradioactivity

All parts have a traceable historythat shows they have low enoughradioactivity to be used. This willbe documented in a database andthere will be an acceptance sheetwith sign off that the part is usable.

Control backgroundlevel in the experi-ment

R-190002 1.9Parts used areclean

All parts have a cleaning procedureto ensure surface contamination isat an acceptable level. This willbe documented in a database andthere will be an acceptance sheetwith sign off that the part is usable.

Control backgroundin the experiment.Control contami-nation of Xenonthat would reduceelectron drift lengthand/or increase lightabsorption

R-190003 1.9Parts are movedand lifted with-out damage

Moving and lifting will be done toprocedures written and approved bysubsystem experts. This will in-clude analysis of rigging attachmentand support loads where applicable.Workers will review the proceduresand have rigging experience

Detector is assem-bled without damag-ing parts

R-190004 1.9Parts are cor-rectly assembled

Assembly work will be done to pro-cedure written and approved bysubsystem experts. Workers will re-view the procedure and have train-ing where necessary.

Detector is assembledcorrectly

R-190005 1.9Control Part ex-posure to Radon

Parts have an allowable total expo-sure to Radon, based on materialand location. Monitored time of ex-posure and Radon level of air arerecorded in a database. For criticalparts, may use samples.

Control backgroundlevel in the experi-ment

337


12.2.10 WBS 1.10 Cleanliness and Screening

The primary requirements for WBS 1.10 Cleanliness and Screening are shown in Table 12.2.10. At Level2, the WBS 1.10 requirements dictate that the project has the sensitivity and capacity to screen all detectorcomponents at the level needed to ensure that the radioactivity requirements are satisfied. WBS 1.10 alsokeeps track of allowed radioactivity levels for individual components, and more details can be found inChapter 9.

Table 12.2.10: Level 2 Requirements for WBS 1.10 Cleanliness and Screening


Rationale

R-200001 1.10Assay Sensitivity for U andTh by Direct Counting

10 pptSensitivity needed to assess to-tal radioactive background

R-200002 1.10Assay Sensitivity for U andTh by Neutron ActivationAnalysis

1.5 pptSensitivity needed to assess to-tal radioactive background

R-200003 1.10Assay Sensitivity for U andTh by ICP-MS

10 pptSensitivity needed to assess to-tal radioactive background

R-200004 1.10Assay Sensitivity for RadonEmanation

0.3mBqSensitivity needed to assess to-tal radioactive background

R-200005 1.10 Assay Sensitivity for210Pb

in the bulk10mBq/kg

Sensitivity needed to assess to-tal radioactive background

R-200006 1.10 Assay Sensitivity for210Pb

Plateout60 nBq/cm2

Sensitivity needed to assess to-tal radioactive background

R-200007 1.10Assay Sensitivity for DustAccumulation 10 ng/cm

2 Sensitivity needed to assess to-tal radioactive background

338


12.2.11 WBS 1.11 Offline Computing

The primary requirements for WBS 1.11 Offline Computing are shown in Table 12.2.11, and are primarilythe amount of disk and CPU to enable adequately detailed studies of the detector and background levels to beperformed. Without sufficient computing resources, the project is unable to determine whether it is satisfyingthe top level science requirements. WBS 1.11 is also critically important in coordinating computing andsoftware tasks with other WBS elements, particularly in simulations and analysis tools.

Table 12.2.11: Level 2 Requirements for WBS 1.11 Offline Computing


Rationale

R-210001 1.11US Disk CostsFY15-FY18

Procure NERSCGPFS disk forsimulation, devel-opment, and initialdata taking

Handle simulation data, deriveddata, and user data volumes as de-fined in Table 11.2.2

R-210002 1.11US CPU CostsFY15-FY18

Procure NERSCPDSF CPU forsimulation, devel-opment, and initialdata taking


R-210003 1.11US Server CostsFY15-FY18

Procure and main-tain servers forsimulation anddevelopment


R-210004 1.11 UK Data Center

Design, deploy,test, and maintainUK Data Centerat Imperial

Handle simulation data, deriveddata, and user data volumes as de-scribed in Sec. 11.2.2

339


12.3 Sensitivity and Detector Performance

We employ the simulation tools described in Section 12.1 to evaluate the sensitivity of LZ, as defined by therequirements in Section 12.2, to a WIMP signal. Additionally, we evaluate the impact of variations in theparameters that define the LZ apparatus on its sensitivity.

The high-statistics calibrations performed by LUX and their incorporation into NEST have driven a con-version from the simple cut-and-count statistical methods used in the CDR to more advanced likelihoodmethods. The use of likelihood methods has markedly reduced the sensitivity of LZ to the dominant electronrecoil backgrounds. Table 12.3.1 shows the number of counts with S1 signals between 0 and 20 photoelec-trons coming from different background sources in a 1,000 d run with no discrimination applied, analogousto Table 9.2.7 (see Sec. 12.1.4 for further discussion of the cuts used to generate these numbers). For themajority of the chapter, we use the PLR method described in Sec. 12.3.1 to exploit the ER/NR discriminationpower of liquid xenon to calculate our sensitivity to dark matter.

Table 12.3.1: Backgrounds described by PDFs in the profile likelihood analysis, with the counts expectedwith S1 signal between 0 and 20 photoelectrons in a 1,000 d run, with no discrimination applied, analogousto Table 9.2.7 (see Sec. 12.1.4 for further discussion of the cuts used to generate these numbers).

Background Type Counts8B solar � NR 7hep � NR 0.21DSN � NR 0.05ATM � NR 0.46pp +7Be +14N solar � ER 255136Xe (2��) ER 6785Kr ER 24.5222Rn ER 722220Rn ER 122Detector components + Environmental ER 11.3Detector components + Environmental NR 0.5

Our evaluation of the LZ sensitivity should be considered as a snapshot in time. The field of direct darkmatter detection with liquid xenon TPCs continues to achieve increasing sensitivity. Since the completionof the LZ CDR [1], the LUX Collaboration has substantially advanced the understanding of the response ofliquid xenon to both background and signal [3, 5], knowledge which has been incorporated in this report. Asthe LZ construction project progresses, key response and radioactive background properties will be measuredand our sensitivity evaluation will adapt. We endeavor to bound this evolution by providing a "baseline"parameter set as well as "goal" and "reduced" sets, summarized in Table 12.3.2. The variation of sensitivitywith excursions in many LZ parameters is captured in Section 12.3.3.

12.3.1 Profile Likelihood Ratio Method

The sensitivity projections in this report are based on a profile likelihood ratio (PLR) method [4], which al-lows near-optimum exploitation of the differences between signal and background in the key parameters thatare reconstructed by the LZ apparatus. The parameters with the highest sensitivity to these differences are theposition-corrected S1 (primary scintillation light) and S2 (secondary luminescence from ionization) signals.The position of events in radius r and height z also allows distinction between signal and the backgroundevents which originate from radioactive impurities in the material in the vicinity of the liquid xenon TPC

340

LZ Technical Design Report 12.3 Sensitivity and Detector Performance

of LZ; these background events cluster near the edges of the TPC, while the WIMP signal is uniform in theliquid xenon mass. In this report, we apply a simple fiducial volume cut, where 5.6 tonnes of the inner liquidxenon is retained as the sensitive volume. In the future, our sensitivity studies will exploit the distinct spatialdistributions of signal and background just as the LUX experiment has done, and the fiducial requirementwill become unnecessary.

0 5 10 15 20 25 30S1c [phd]

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

log1

0(S

2c/

S1

c)

8B

ER

Figure 12.3.1: Simulations of the most prominent ER and NR (from 8B) backgrounds are plotted inthe log10(S2c∕S1c)-S1c plane. The statistics shown represent 5x the expected ER background and 500xthe expected NR background in the nominal LZ exposure. The red tinted area shows the expectation forevents from a 40 GeV/c2-mass WIMP, falling between the two background populations with the regionenclosed by the solid(dashed) line representing the 1 �(2 �) band.

To execute the PLR sensitivity estimate, signal and background probability distribution functions (PDFs)are created in S1 and S2 after application of the fiducial cut in r and z. The signal PDF for each WIMP massis generated by converting the differential energy spectrum calculated from [20] to S1 and S2 signals in theLZ detector using NEST and the parameterization of detector response described in Section 12.1.3.

The background PDFs are broken into the eleven individual components listed in Table 12.3.1. The sim-ulations for detector components and environmental backgrounds are summed together into a single PDFeach for ER and NR events. For each WIMP mass, we scan over the cross section to set a 90 % confidenceinterval (CI) for the expected number of signal events, evaluated using RooStats [21]. In the PLR techniquewe use the unbinned likelihood computed in the plane of log10(S2∕S1) versus S1. Poisson fluctuations areinnate to the technique.

The power of the PLR technique arises from an optimal weighting of the background-free and background-rich regions in the log10(S2∕S1)-S1 plane. Figure 12.3.1 shows high-statistics simulations of the most promi-nent backgrounds (ER events from pp solar neutrinos, 222Rn, and 220Rn, and NR events from 8B neutrinos)in the log10(S2∕S1)-S1 plane, representing 5x and 500x the count rates expected in the nominal LZ exposurefor ER and NR, respectively. Also shown is the region that would be populated by events from a 40 GeV/c2

WIMP signal, which falls between the 8B and ER background areas. The PLR technique optimally combinesthe background-free and background-rich regions. The intense calibrations recently conducted by the LUXcollaboration provide confidence in the knowledge of the shapes of the backgrounds and the signal.

341


10 5 0 5 10log10(L(BG)/L(Signal))

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08N

orm

aliz

ed C

ounts

40 GeV WIMP Signal

Background

Figure 12.3.2: A one-dimensional projection of the PLR discrimination statistic. Two ensembles ofpoints in the log10(S2∕S1)-S1 plane are considered, one distributed like a 40 GeV/c

2 WIMP signal, andthe other like the expected background (combining both the ER and 8B bands of Fig. 12.3.1).

0 5 10 15 20 25 30S1c [phd]

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

log10(S

2c/

S1c)

8B

ER

0 5 10 15 20 25 30S1c [phd]

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

log10(S

2c/

S1c)

8B

ER

Figure 12.3.3: PLR technique for different masses. The background distributions are the same as inFigure 12.3.1, but the expected signal regions are for a 10 GeV/c2 WIMP (left) and a 1,000GeV/c2

WIMP (right). The signal regions merge, respectively, into the 8B and ER background regions. Theexpected signal regions are tinted red, with the darker(lighter) color 1 �(2 �).

342


The discrimination statistic that quantifies whether each point in the respective ensemble more resemblesbackground or signal is evaluated in one dimension as the difference between logarithms of the likelihoodsthat a point in the ensemble is background or signal. Low values correspond to poor likelihood to be back-ground, and high values to a good likelihood to be background. Figure 12.3.2 shows the PLR discriminationstatistic integrated over the full energy range for a 40 GeV/c2 WIMP signal and the expected background(combining both the ER and 8B bands of Fig. 12.3.1). Comparison of the two ensembles shows the consid-erable separation between signal and background available using this method. This plot and similar plotsconstructed for different WIMP masses are used to describe the discrimination power of the LZ apparatusagainst background.

99.99% 99% 90%Background Rejection

0.1

0.2

0.3

0.4

0.6

0.7

0.8

0.9

1

Sig

nal A

ccepta

nce

10 GeV

15 GeV

20 GeV

25 GeV

30 GeV

40 GeV

50 GeV

100 GeV

1000 GeV

99.9%

0.5

Figure 12.3.4: Acceptance and rejection for WIMP signals in LZ. For a variety of WIMP masses,histograms like that shown in the Fig. 12.3.2 are integrated to derive the curves shown. Backgroundsfrom both 8B and ER events are included. The requirement of 99.5% rejection at 50% acceptance isprojected for all WIMP masses.

The expected signal region varies according to the WIMP mass, and Figure 12.3.3 shows the signal re-gions for WIMP masses of 10 GeV/c2 and 1,000 GeV/c2 compared to the same ER and NR simulations ofFig. 12.3.1. The PLR method naturally takes into account the shape of the expected WIMP signal, and theachievable discrimination at all WIMP masses is thus significantly better than that attainable in the ‘cut andcount’ technique utilized in the LZ CDR. Figure 12.3.4 shows the signal acceptance for WIMPs of a varietyof masses as a function of the fraction of the background rejected. For all WIMP masses, a backgroundrejection that exceeds 99.5 % for signal acceptance of 50 % is projected.

12.3.2 LZ Sensitivity Projection

To evaluate the projected sensitivity for LZ, we consider a run of 1,000 live days and a 5.6 tonne fiducialmass. We assume the same background models of Table 12.3.1, where background counts were shown for

343


a signal region encompassing 0 to 20 phd, effectively the search region for 100 GeV/c2 WIMP masses andbelow. To increase the sensitivity over the entire WIMP mass range, we consider an expanded S1 signalregion of 0 to 50 phd, corresponding to 1.5 to 16 keVee for ER and 6 to 60 keVnr for NR. Although theexpanded search region brings with it a higher absolute count of backgrounds than those listed in the table,the use of the PLR method smoothly accounts for the profiles of these backgrounds relative to each WIMPmass as described in the previous section. We include the requirement of a 3-fold coincidence for the PMTs.

]2WIMP mass [GeV/c10 210 310

) [p

b]S

Ipσ(

10lo

g

-14

-13

-12

-11

-10

-9

-8

-7

-6

-5

-4

-3

]2 [c

mS

Ipσ

-5010

-4910

-4810

-4710

-4610

-4510

-4410

-4310

-4210

-4110

-4010

-3910

-14

-13

-12

-11

-10

-9

-8

-7

-6

-5

-4

-3

10 210 310

scattering-N coherentν

1 event

1000 tonne-years

significance

σ3

LZ projected

90% CL Median CDR

90% CL Median (Baseline)

90% CL Median (Goal)

Zeplin-III (2012)

LUX WS2013+WS2014-16

Figure 12.3.5: LZ sensitivity projection. The baseline LZ assumptions in this report give the solid blackcurve. LUX and ZEPLIN results are shown in broken blue lines. If LZ achieves the design goals listedin Table 12.3.2, the sensitivity would improve, resulting in the pink sensitivity curve. The green lineshows the projected sensitivity in the LZ Conceptual Design Report (CDR) [1] (see text for details ofthe changes from the CDR to this report). Lastly, the shaded regions show where coherent scatteringneutrino backgrounds emerge.

The projected sensitivity curve for LZ is shown in Figure 12.3.5. The best sensitivity is 2.3 × 10−48 cm2

for a 40 GeV/c2 WIMP mass, which satisfies our top level science requirement.Figure 12.3.5 also shows the projected sensitivity from the LZ Conceptual Design Report (CDR) [1].

There are three main differences in this projection with respect to the CDR:

1. We have increased the assumed level of 222Rn and 220Rn by a factor of twenty, so radon is now thedominant source of ER events in the detector.

2. The PLR technique as described in the previous section is used to evaluate the dark matter sensitivity,instead of a simple cut and count approach. These two changes effectively offset: there is a higher

344


overall background level but those backgrounds are more effectively rejected by the analysis technique.Our confidence in the ER background shape, most important for using the PLR technique, rests on thehigh-statistics LUX tritium ER calibration [2].

3. The LUX neutron calibration results, which allow LZ to more confidently project the response ofthe detector to very low energy nuclear recoils. Use of the neutron calibration provides the greatlyincreased sensitivity to very low-mass WIMPs in the new projection.

Figure 12.3.6 shows the discovery potential for LZ under the baseline assumptions, calculated using acut-and-count technique via the TRolke package [22, 23]. With the baseline parameters, LZ will have 3�significance for 40 GeV/c2 WIMP mass at a cross section of 6.0 × 10−48 cm2. We also show the 5� sig-nificance curve, which falls below the projections from the XENON1T experiment at all WIMP masses.Figure 12.3.7 shows the background and signal events populating the log10(S2∕S1)-S1 plane in an exampleLZ experiment with 1000 days and 5600 kg exposure for the 3� example combination of WIMP parameters.


) [p

b]S

Ipσ(

10lo

g

-12

-10

-8

]2 [c

mS

Ipσ

-4910

-4810

-4710

-4610

-4510

-4410

-4310

scattering-N coherentν 1 event

1000 tonne-years

significance

σ3

90%CL X

ENON1T

LZ projected

Median Significanceσ5

Median Significanceσ3

90% CL Median (Baseline)

Figure 12.3.6: The discovery potential for LZ under the baseline assumptions, calculated using a cut-and-count technique via the TRolke package [22, 23]. With the baseline parameters, LZ will have3� significance for 40 GeV/c2 WIMP mass at a cross section of 6.0 × 10−48 cm2. The 5� significanceexpectation is just below the expected 90 % CL limit from a two year run of XENON1T at all WIMPmasses.

12.3.3 Parameter Scans of LZ Sensitivity

We have explored the dependency of the LZ sensitivity to spin-independent interactions of WIMPs uponcritical detector performance assumptions. Our baseline parameters lead to the sensitivity shown by thesolid blue curve in Fig. 12.3.5

345


0 5 10 15 20 25 30S1c [phd]

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

3.8

4.0

log10(S

2c/

S1c)

1 keVee5 keVnr

3 keVee14 keVnr

5 keVee22 keVnr

ER Background

WIMP (3σ significance, 40 GeV)8B Background

Figure 12.3.7: A sample LZ exposure under the baseline assumptions where the signal represents a40GeV/c2 WIMP mass with a cross section of 6.0 × 10−48 cm2. LZ will observe several 8B events underthese assumptions, along with the nominal ER backgrounds. For these WIMP parameters, LZ will havea 3� median significance for WIMP discovery. The solid blue and red lines represent the median of theER and NR bands, the dashed blue and red lines indicate the 10 % to 90% intervals for each population.The dashed lines running from the top left corner down to the x-axis show lines of constant recoil energy.

One such study defines “goals” as a set of parameters that are likely to be achieved during the fabricationof LZ, but which are somewhat less conservative than the baseline. We define a "reduced" set of parametersas an unlikely worst-case. Both the goal and baseline parameter sets meet all requirements in Section 12.2.These three parameters sets are listed in Table 12.3.2, and their projected performance is shown in Fig-ure 12.3.8. Reaching the LZ goals achieves a sensitivity of 1.1 × 10−48 cm2 at 40 GeV/c2. The reducedparameter set degrades the sensitivity to 5.1 × 10−48 cm2 at 40 GeV/c2.

Table 12.3.2: Key parameters for reduced, baseline and goal detector performance as explained in thetext.

Detector Parameter Reduced Baseline GoalLight collection (PDE) 0.05 0.075 0.12Drift field (V/cm) 160 310 650Electron lifetime (µs) 850 850 2800PMT phe detection 0.8 0.9 1.0N-fold trigger coincidence 4 3 2222Rn (mBq in active region) 13.4 13.4 0.67Live days 1000 1000 1000

346


A notable observation that would accompany the achievement of the “goal” set of parameters: over 300 8Bneutrino events would be observed in a 1,000 live days LZ run. While these events would slow the discoveryof a WIMP in the 7 GeV/c2 mass range, they would also demonstrate a physical process not yet observed innature, the coherent scattering of neutrinos from nuclei.


) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410LZ 90% CL

Reduced

Baseline

Goal

Figure 12.3.8: LZ sensitivity projections for goal, baseline, and reduced parameters. The respective bestsensitivities are 1.1 × 10−48 cm2, 2.3 × 10−48 cm2 (which both satisfy the primary science requirement),and 5.1 × 10−48 cm2 at a WIMP mass of 40GeV/c2.

In the next several figures, we vary key detector performance parameters to gauge their impact on themedian 90 % CL of the upper limit for WIMP cross-section. Figure 12.3.9 shows the impact of 222Rn, thedominant ER background, on LZ sensitivity. Assuming that the goal of 0.67 mBq is achieved, the sensitivity(to a 40 GeV/c2 WIMP) improves by 20 % over the baseline case. If the Rn background is increased another10x over the nominal case then the sensitivity would degrade 20 %. The 222Rn rate is representative of ageneric flat ER background in LZ. Even a factor of 10 increase in the radon has marginal impact on sensitivity,demonstrating the power of using the PLR to separate ER backgrounds from signal like events. The effectof radon on the discovery potential is more significant, as shown in Fig. 12.3.10, where a 10x increase in theradon rate degrades the discovery potential of the detector by a factor of two.

Figure 12.3.11 shows the impact on sensitivity of scaling the atmospheric (ATM) coherent scatteringneutrino background. ATM neutrinos produce nuclear recoils similar to that of a 40 GeV/c2 WIMP, andtheir PDF has a high degree of overlap with the WIMP signal. Adjusting the ATM rate serves as a proxy forchanging the overall NR count. When the ATM rate is turned up by a factor of 10, it is equivalent to havingroughly 5 extra NR counts, degrading the sensitivity (to a 40 GeV/c2 WIMP) by 25 %.

Figure 12.3.12 shows the effect of changes in the S1 photon detection efficiency alone. Greater lightcollection leads to better S1 resolution, tightening the distribution of NR and ER events and leading toimproved discrimination. Better light collection also improves the low energy threshold of LZ, enhancingsensitivity to low-mass dark matter, although this effect is somewhat countered by also seeing a higher 8Bbackground at low energy. Figure 12.3.13 shows the effect of changes in the purity of the LXe, as represented

347


by the electron lifetime. There is significant margin until the drift time drops below half the baseline value.Figure 12.3.14 gives the dependence on electron extraction efficiency. With only 50 % extraction efficiency,larger fluctuations in the S2 signal lead to a reduction in discrimination power and a corresponding loss ofsensitivity. Figure 12.3.15 shows that the sensitivity depends weakly on the drift field. Figure 12.3.16 showsthe dependence of the sensitivity on the coincidence trigger requirement, where the baseline design assumesa 3-PMT coincidence trigger. Going to 2-PMT coincidence reduces the threshold and makes a significantimpact for low-mass WIMPs and similarly the 8B neutrino signal.

Lastly, Figure 12.3.17 shows how extending the run from 1,000 to 3,000 live days would improve thesensitivity of the experiment. The plot shows the median 90 % confidence level upper limit on the crosssection for a 40 GeV/c2 WIMP. In the baseline case, the sensitivity can be improved from 2.3 × 10−48 cm2 to1.3 × 10−48 cm2. If all the design goals are achieved, the sensitivity can be improved from 1.1 × 10−48 cm2

to 6 × 10−49 cm2 with 3,000 live days. Figure 12.3.17 also shows the 1� bands on the expected sensitivityfor both the baseline and goal parameter sets in a given exposure of LZ.


) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410Rn-222

5x, 67 mBq

Baseline, 13.4 mBq

Goal, 0.67 mBq

Figure 12.3.9: LZ sensitivity projections for three different assumptions on the concentration of radonin the active volume.

348



) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410 median significanceσRn-222, 3

10x, 134 mBq

Baseline, 13.4 mBq

Goal, 0.67 mBq

Figure 12.3.10: LZ 3� median significance projections for three different assumptions on the concen-tration of radon in the active volume.


) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8]2

[cm

SI

pσ

-4810

-4710

-4610

-4510

-4410νATM

50x, 23 counts

10x, 4.6 counts

Baseline, 0.46 counts

Figure 12.3.11: LZ sensitivity projections vs. scaled atmospheric neutrino flux. Scaling the atmosphericneutrino rate is a proxy for scaling the overall NR backgrounds in LZ.

349



) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410Photon detection efficiency

0.05

0.075 (Baseline)

0.12

Figure 12.3.12: LZ sensitivity projections vs. S1 photon detection efficiency. We assume a photondetection efficiency of 0.075 for the baseline sensitivity, matching the requirement. The current modelof the detector predicts a value of 0.085 (see Sec. 3.5.1).


) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8]2

[cm

SI

pσ

-4810

-4710

-4610

-4510

-4410Electron lifetime

sµ500

s (Baseline)µ850

sµ2800

Figure 12.3.13: LZ sensitivity projections vs. electron lifetime (for the nominal electric field value of310 V/cm). LZ does not lose significant sensitivity until the lifetime drops below half the nominal valueof 850 µs.

350



) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410Electron extraction efficiency

50%

95% (Baseline)

99%

Figure 12.3.14: LZ sensitivity projections vs. electron extraction efficiency. With only 50 % extrac-tion efficiency, larger fluctuations in the S2 signal lead to a reduction in discrimination power and acorresponding loss of sensitivity.


) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8]2

[cm

SI

pσ

-4810

-4710

-4610

-4510

-4410Drift Field

650 V/cm310 V/cm (Baseline)160 V/cm50 V/cm

Figure 12.3.15: LZ sensitivity projections vs. electric field. In this regime, the ER/NR discrimination isrobust to changes in the drift field, and the effect on sensitivity is minor.

351



) [p

b]S

Ipσ(

10lo

g

-12

-11

-10

-9

-8

]2 [c

mS

Ipσ

-4810

-4710

-4610

-4510

-4410N-fold PMT coincidence

4-fold

3-fold (Baseline)

2-fold

Figure 12.3.16: LZ sensitivity projections vs. trigger coincidence level. The primary effect of thecoincidence requirement is on the detection of very low energy events, with direct consequences forsensitivity to low WIMP masses. For comparison, the LUX detector operates with a 2-fold triggercoincidence.

Live time [days]0 1000 2000 3000

2] a

t 40

GeV

/c2

[cm

SI

pσ

-4910

-4810

-4710

90% CL Median (Baseline) (sys.)σ1±

90% CL Median (Goal) (sys.)σ1±

Figure 12.3.17: LZ sensitivity projections vs. exposure for the baseline (blue) and goal (orange).The impact of running the experiment for an additional 2000 days is to improve the sensitivity to1.2 × 10−48 cm2 and 6 × 10−49 cm2 for the baseline and goal scenarios, respectively. Also shown are the1� bands showing the range of possible sensitivities for a given exposure.

352

LZ Technical Design Report 12.4 Bibliography

12.4 Bibliography

[1] D. S. Akerib et al. (LZ), (2015), Conceptual Design Report; LBNL-190005, FERMILAB-TM-2621-AE-E-PPD, arXiv:1509.02910 [physics.ins-det].

[2] D. S. Akerib et al. (LUX), Phys. Rev. D93, 072009 (2016), arXiv:1512.03133 [physics.ins-det].

[3] D. S. Akerib et al. (LUX), “Low-energy (0.7-74 keV) nuclear recoil calibration of the LUX darkmatter experiment using D-D neutron scattering kinematics,” (2016), submitted to Phys. Rev. C,arXiv:1608.05381 [physics.ins-det].

[4] G. Cowan, K. Cranmer, E. Gross, and O. Vitells, Eur. Phys. J. C71, 1554 (2011), [Erratum: Eur. Phys.J.C73,2501(2013)], arXiv:1007.1727 [physics.data-an].

[5] D. S. Akerib et al. (LUX), Phys. Rev. Lett. 116, 161301 (2016), arXiv:1512.03506 [astro-ph].

[6] D. S. Akerib et al. (LUX), Nucl. Instrum. Meth. A675, 63 (2012), arXiv:1111.2074 [physics.data-an].

[7] S. Agostinelli et al. (GEANT4), Nucl. Instrum. Meth. A506, 250 (2003).

[8] M. Szydagis, N. Barry, K. Kazkaz, J. Mock, D. Stolp, M. Sweany, M. Tripathi, S. Uvarov, N. Walsh,and M. Woods (NEST), J. Instrum. 6, P10002 (2011), arXiv:1106.1613 [physics.ins-det].

[9] M. Szydagis, A. Fyhrie, D. Thorngren, and M. Tripathi (NEST), Proceedings, LIght Detection InNoble Elements (LIDINE2013), J. Instrum. 8, C10003 (2013), arXiv:1307.6601 [physics.ins-det].

[10] W. B. Wilson et al., SOURCES4A: A Code for Calculating (alpha,n), Spontaneous Fission, and DelayedNeutron Sources and Spectra, Tech. Rep. LA-13639-MS (Los Alamos, 1999).

[11] M. J. Carson et al., Astropart. Phys. 21, 667 (2004), arXiv:hep-ex/0404042 [hep-ex].

[12] R. Lemrani, M. Robinson, V. A. Kudryavtsev, M. De Jesus, G. Gerbier, and N. J. C. Spooner, Nucl.Instrum. Meth. A560, 454 (2006), arXiv:hep-ex/0601030 [hep-ex].

[13] V. Tomasello, M. Robinson, and V. A. Kudryavtsev, Astropart. Phys. 34, 70 (2010).

[14] M. Herman, R. Capote, B. Carlson, P. Oblozinsky, M. Sin, A. Trkov, H. Wienke, and V. Zerkin, Nucl.Data Sheets 108, 2655 (2007).

[15] V. A. Kudryavtsev, Comput. Phys. Commun. 180, 339 (2009), (MUSIC), arXiv:0810.4635[physics.comp-ph].

[16] P. Antonioli, C. Ghetti, E. V. Korolkova, V. A. Kudryavtsev, and G. Sartorelli, Astropart. Phys. 7, 357(1997), (MUSIC), arXiv:hep-ph/9705408 [hep-ph].

[17] B. Lenardo, K. Kazkaz, A. Manalaysay, J. Mock, M. Szydagis, and M. Tripathi (NEST), IEEE Trans.Nucl. Sci. 62, 3387 (2015), arXiv:1412.4417 [astro-ph.IM].

[18] C. E. Dahl, The physics of background discrimination in liquid xenon, and first results from Xenon10in the hunt for WIMP dark matter, Ph.D. thesis, Princeton U. (2009).

[19] C. H. Faham, V. M. Gehman, A. Currie, A. Dobi, P. Sorensen, and R. J. Gaitskell, J. Instrum. 10,P09010 (2015), arXiv:1506.08748 [physics.ins-det].

[20] C. McCabe, Phys. Rev. D82, 023530 (2010), arXiv:1005.0579 [hep-ph].

353

http://lz.lbl.govhttp://arxiv.org/abs/1509.02910http://luxdarkmatter.orghttp://dx.doi.org/ 10.1103/PhysRevD.93.072009http://arxiv.org/abs/1512.03133http://luxdarkmatter.orghttp://arxiv.org/abs/1608.05381http://dx.doi.org/ 10.1140/epjc/s10052-011-1554-0, 10.1140/epjc/s10052-013-2501-zhttp://arxiv.org/abs/1007.1727http://luxdarkmatter.orghttp://dx.doi.org/ 10.1103/PhysRevLett.116.161301http://arxiv.org/abs/1512.03506http://luxdarkmatter.orghttp://dx.doi.org/ 10.1016/j.nima.2012.02.010http://arxiv.org/abs/1111.2074http://www.geant4.org/geant4/http://dx.doi.org/ 10.1016/S0168-9002(03)01368-8http://nest.physics.ucdavis.edu/site/http://dx.doi.org/ 10.1088/1748-0221/6/10/P10002http://arxiv.org/abs/1106.1613http://nest.physics.ucdavis.edu/site/http://dx.doi.org/ 10.1088/1748-0221/8/10/C10003http://arxiv.org/abs/1307.6601https://www.osti.gov/servlets/purl/15215-mGfIQS/webviewable/https://www.osti.gov/servlets/purl/15215-mGfIQS/webviewable/http://dx.doi.org/ 10.1016/j.astropartphys.2004.05.001http://arxiv.org/abs/hep-ex/0404042http://dx.doi.org/10.1016/j.nima.2005.12.238http://dx.doi.org/10.1016/j.nima.2005.12.238http://arxiv.org/abs/hep-ex/0601030http://dx.doi.org/10.1016/j.astropartphys.2010.05.005http://dx.doi.org/10.1016/j.nds.2007.11.003http://dx.doi.org/10.1016/j.nds.2007.11.003http://dx.doi.org/10.1016/j.cpc.2008.10.013http://arxiv.org/abs/0810.4635http://arxiv.org/abs/0810.4635http://dx.doi.org/ 10.1016/S0927-6505(97)00035-2http://dx.doi.org/ 10.1016/S0927-6505(97)00035-2http://arxiv.org/abs/hep-ph/9705408http://nest.physics.ucdavis.edu/site/http://dx.doi.org/ 10.1109/TNS.2015.2481322http://dx.doi.org/ 10.1109/TNS.2015.2481322http://arxiv.org/abs/1412.4417http://search.proquest.com/docview/304986842http://search.proquest.com/docview/304986842http://dx.doi.org/ 10.1088/1748-0221/10/09/P09010http://dx.doi.org/ 10.1088/1748-0221/10/09/P09010http://arxiv.org/abs/1506.08748http://dx.doi.org/10.1103/PhysRevD.82.023530http://arxiv.org/abs/1005.0579


[21] L. Moneta, K. Belasco, K. S. Cranmer, S. Kreiss, A. Lazzaro, D. Piparo, G. Schott, W. Verkerke, andM. Wolf, Proceedings, 13th International Workshop on Advanced computing and analysis techniquesin physics research (ACAT2010), PoS ACAT2010, 057 (2010), arXiv:1009.1003 [physics.data-an].

[22] J. Lundberg, J. Conrad, W. Rolke, and A. Lopez, Comput. Phys. Commun. 181, 683 (2010),arXiv:0907.3450 [physics.data-an].

[23] W. A. Rolke, A. M. Lopez, and J. Conrad, Nucl. Instrum. Meth. A551, 493 (2005),arXiv:physics/0403059 [physics].

354
http://arxiv.org/abs/1009.1003http://dx.doi.org/ 10.1016/j.cpc.2009.11.001http://arxiv.org/abs/0907.3450http://dx.doi.org/10.1016/j.nima.2005.05.068http://arxiv.org/abs/physics/0403059

Documents

12 Simulations, Requirements, and Detector Performancehep.ucsb.edu/LZ/TDR/12_Simulations_Requirements_and... · 2017-03-26 · Figure 12.1.1: Engineering drawing and simulation geometry